UV Absorption in NGC 5548 Jerry Kriss STScI 8/17/2017
The Narrow Absorption Components in NGC 5548 This is the C IV region prior to the 2013 XMM campaign. Absorption components are labeled from high to low velocity.
The Bottoms of the Narrow C IV Absorption Troughs in NGC 5548 Do Not Vary
Changes in the X-ray Spectrum of NGC 5548 As of December 2016, Strong Soft X-ray Absorption is Still Obscuring NGC 5548 Kaastra+14
Broad C IV Absorption in NGC 5548 Kaastra+14
Normalized Broad Absorption Profiles in NGC 5548 Kaastra+14 Kaastra+14
The Obscurer in NGC 5548 also Shadows the Warm Absorbers Kaastra+14
The X-ray Obscurer also Obscures the UV-Ionizing part of the Spectral Energy Distribution Arav+15
The Obscured Ionizing Continuum of NGC 5548 Reveals Low-ionization Narrow Absorption Lines In addition to Si III, Si II, C III, C II, and PV absorption lines are also present in Component #1. Arav+15
Photoionization Solutions for Component #1 (all 5 epochs) Arav+15
« Density-sensitive lines in Component #1 give: log n e = 4.8±0.1 « For log(U)= − 1.5, distance R=3 pc « This is comparable to the NLR size of 1—3 pc (Peterson et al. 2013) Arav+15
Using Recombination Times to Measure Density « Following Krolik & Kriss (1995), a simple model for time-dependent photoionization effects is given by dn i /dt = − (F i σ ion,i + n e α rec,i − 1 )n i + n e n i+1 α rec,i + F i − 1 σ ion,i − 1 n i − 1 « For the ions we will be measuring, generally n i-1 << n i << n i+1 , so dn i /dt = − F i σ ion,i n i + n e n i+1 α rec,i « In general, as long as there is a copious increase in ionizing flux, the − F i σ ion,i n i b term dominates, and ions n i are destroyed instantly. Conversely, when the flux decreases dramatically, n e n i+1 α rec,i dominates, and n i reappears more slowly.
F(1367) Recombination Ionization
To measure Time delays better, Use only intervals of decreasing flux: F(1367)
Atomic Physics Really Works! Ionization/recombination in the absorption lines smear and delay their response. These time delays also give densities (log): C II #1 5.0 ± 0.3 cm − 3 Si III #1 5.3 ± 0.3 cm − 3 Si IV #1 4.8 ± 0.2 cm − 3 Si III #3 5.1 ± 0.3 cm − 3 Si IV #3 4.8 ± 0.2 cm − 3 Densities are consistent with C III* and Si III* for Component #1. Kriss+ in prep Distances again are 3 − 5 pc, Kriss+ in prep the same as the NLR of NGC 5548.
Possibilities for the BLR Holiday a) Part of the BLR (a large part) is being shadowed by some structure that blocks the continuum to the BLR but not our line of sight. b) The continuum itself has changed, so that the SED has changed; given the depths of the decorrelation, it must be the high-energy portion of the ionizing continuum that has changed the most (e.g., He II decorrelates with the 1367 Å flux by 21%, Ly α by only 9%).
Peculiar responses of the Narrow Absorption Lines - Facts « The narrow absorption lines respond to continuum variations. « Throughout the first two continuum peaks and troughs, the response is classic, textbook photoionization (instantaneous) and recombination (delayed, and density dependent). « During the second half of the campaign, particularly throughout the BLR holiday, the absorption lines also decorrelate from the continuum. Peculiarly, some do not . . . C II λ 1334, and Ly α . . . « The following slides show light curves for narrow absorption by Component #1 ordered by increasing ionization potential: Ly α λ 1216 (13.6 eV), Si II λ 1526 (16.3 eV), C II λ 1334 (24.4 eV), Si III λ 1206 (33.5 eV), Si IV λ 1393 and λ 1402 (45.1 eV), C III* λ 1775 (47.9 eV), C IV λ 1548 (64.5 eV), and N V λ 1238 (97.9 eV).
13.6 eV F(1367) Note the good correlation here!
F(1367) 16.3 eV Some correlation
F(1367) 24.4 eV Some correlation
F(1367) 33.5 eV NO correlation!
F(1367) 45.1 eV
47.9 eV
64.5 eV
97.7 eV
This change must be affecting the ionizing continuum at energies >~30 eV. SiII NV SiIII Arav+15
« This change would affect the high-energy tail of the warm, Comptonized inner region of the disk « Is it a change in the intrinsic flux, or a change in the obscurer? 940 Å (14&16 Jan 2016)
Inferring the SED in the Extreme Ultraviolet « As Gary Ferland noted last summer, “The absorption lines are the key. They’re along the line of sight, and they see the same continuum.” « Take Gary at his word. To a good first approximation, the ionizing flux at a line’s ionization potential IS the ionizing flux driving that line (or, is at least directly proportional to it). « For the first 55 days of the campaign, this ionizing flux is well tracked by the far-UV continuum flux at 1367 Å seen by HST. « For each line, use this proportionality to derive a relationship between a line’s EW and the continuum flux: EW = a0 + a1 * F1367
Good Correlation for C IV for the First 55 Days
Poor Correlation between EW and F(1367) in C IV for the Full Campaign Anomaly No anomaly
Better Correlation of CIV with the Soft X-ray Flux Anomaly No anomaly
Poor Correlation for NV and F(1367) Anomaly No anomaly
Better Correlation for NV and SX Anomaly No anomaly
Poor Correlation between EW and SX in Ly α Anomaly No anomaly
Better Correlation for Ly α and F(1367) Anomaly No anomaly
Poor Correlation for C II and SX Anomaly No anomaly
Better Correlation of CII with F(1367) Anomaly Pre anomaly Post anomaly
Inferring the SED in the Extreme Ultraviolet « For the first 55 days, when everything correlates well, assume that the continuum shape is that shown in Figure 4 of Mehdipour et al. (2015). Call points on this curve “F0(1367)”, or “F0(IP)”.
NGC 5548 Spectral Energy Distribution 940 Å (14&16 Jan 2016)
Inferring the SED in the Extreme Ultraviolet « For the first 55 days, when everything correlates well, assume that the continuum shape is that shown in Figure 4 of Mehdipour et al. (2015). Call points on this curve “F0(1367)”, or “F0(IP)”. « Use the EW/F1367 correlation to then derive the proportionality between the line EW and the flux at its ionization potential: Flux(IP) = F0(IP)/F0(1367) * (EW − a0)/a1 « We can then derive an ionizing continuum light curve based on the EW light curve of each absorption line for their individual ionization potentials.
EUV Light Curves at the Ionization Potentials of NV, C IV, and Ly α
Comparison of Derived EUV Light Curves to the %difference for C IV (Fig. 1 of Goad et al. 2016)
Soft X-ray Light Curve and the Anomaly (Mathur et al. 2017)
Implications for the BLR Holiday « Changing the ionizing SED is a desirable solution since it can explain at least two phenomena with one effect that does not require a special geometrical arrangement. « Does suppressing the ionizing continuum above 30 eV have the desired effect on the relative response of the broad lines? For example, Paper IV says Ly α is suppressed by 9%, H β by ~30%, and C IV, He II and Si IV by 18-23%. Is this consistent with such a change in shape of the SED? « Still to evaluate: Can simply changing the SED give the observed velocity dependence of the holiday onset and impact? (Need to evaluate this in the context of the gas distribution inferred from Anna and Keith’s models of the BLR.)
The Broad UV Absorption Lines Also Vary with Time Kriss+ in prep
Light Curves for the Broad Absorption Features Kriss+ in prep
Broad Absorption Compared to BLR “Obscuration” (BLR “Holiday”, Paper IV) UV Broad Absorption is at a minimum here Obscuration of the BLR is at a maximum here
Conclusions « Decorrelations in the equivalent widths of the narrow absorption lines of NGC 5548 are associated with the ionization potentials of the absorbing ions. « Lines with higher ionization potentials decorrelate more at times associated with the BLR holiday. « Using the absorption line strengths, we can infer the continuum flux at the line’s ionization potential as a function of time. « Light curves for the inferred extreme ultraviolet continuum flux in the 13.6—97.9 eV range show high-energy diminutions strikingly similar to the flux deficits observed in the broad emission lines during the “BLR Holiday”. « The cause of the BLR Holiday is not obscuration of the continuum source, but a change in the SED (at energies not visible to us).
Recommend
More recommend
